U N I V E R S I T Y O F O U L U P. O. B . 7 5 0 0 F I - 9 0 0 1 4 U N I V E R S I T Y O F O U L U F I N L A N D
U N I V E R S I TAT I S
S E R I E S
SCIENTIAE RERUM NATURALIUM
Professor Mikko Siponen
University Lecturer Elise Kärkkäinen
U N V E R S T AT S O U L U E N S S
U N IIV E R S IIT AT IIS O U L U E N S IIS
E D I T O R S
O U L U E N S I S
A C TA
Professor Hannu Heusala
Professor Olli Vuolteenaho
SCIENTIAE RERUM SOCIALIUM
Senior Researcher Eila Estola
Information officer Tiina Pistokoski
University Lecturer Seppo Eriksson
EDITOR IN CHIEF
Professor Olli Vuolteenaho
Publications Editor Kirsti Nurkkala
ISBN 978-951-42-8905-7 (Paperback)
ISBN 978-951-42-8906-4 (PDF)
ISSN 0355-3213 (Print)
ISSN 1796-2226 (Online)
FACULTY OF TECHNOLOGY,
DEPARTMENT OF PROCESS AND ENVIRONMENTAL ENGINEERING,
UNIVERSITY OF OULU
ACTA UNIVERSITATIS OULUENSIS
C Te c h n i c a 3 0 5
Academic dissertation to be presented, with the assent of
the Faculty of Technology of the University of Oulu, for
public defence in Auditorium TA105, Linnanmaa, on
October 31st, 2008, at 12 noon
O U L U N Y L I O P I S TO, O U L U 2 0 0 8
Illikainen, Mirja, Mechanisms of thermomechanical pulp refining
Faculty of Technology, Department of Process and Environmental Engineering, University of
Oulu, P.O.Box 4300, FI-90014 University of Oulu, Finland
Acta Univ. Oul. C 305, 2008
The objective of this thesis was to obtain new information about mechanisms of
thermomechanical pulp refining in the inner area of a refiner disc gap by studying inter-fibre
refining and by calculating the distribution of energy consumption in the refiner disc gap.
The energy consumption of thermomechanical pulping process is very high although
theoretically a small amount of energy is needed to create new fibre surfaces. Mechanisms of
refining have been widely studied in order to understand the high energy consumption of the
process, however, phenomena in the inner area of disc gap has had less attention. It is likely that
this important position is causing high energy consumption due to the high residence time of pulp
The power distribution as a function of the refiner disc gap was calculated in this work. The
calculation was based on mass and energy balances, as well as temperature and consistency
profiles determined by mill trials. The power distribution was found to be dependent on segment
geometry and the refining stage. However, in the first stage refiner with standard refiner segments,
a notable amount of power was consumed in the inner area of the disc gap.
Fibre-to-fibre refining is likely to be the most important mechanism in the inner area of disc
gap from the point of view of energy consumption. In this work the inter-fibre refining was studied
using equipment for shear and compression. Fibre-to-fibre refining was found to be an effective
way to refine fibres from coarse pulp to separated, fibrillated and peeled fibres if frictional forces
inside the compressed pulp were high enough. It was proposed that high energy of today’s
thermomechanical pulping process could derive from too low frictional forces that heated pulp and
evaporated water without any changes in fibre structure.
The method to calculate power distribution and results of fibre-to-fibre refining experiments
may give ideas for developing today’s thermomechanical pulp refiners’ or for developing totally
new energy saving mechanical pulping processes.
Keywords: disruptive shear stress, energy balance, energy consumption, fibre
development, fibre-to-fibre refining, mass balance, power distribution, pulp pad
refining, refining, thermomechanical pulping, TMP
The research work reported in this thesis was carried out at the Fibre and Particle
Engineering Laboratory, University of Oulu during the period 2003–2007.
Financial support from the Graduate School of Chemical Engineering (GSCE),
UPM-Kymmene and Metso Paper is gratefully acknowledged. I also appreciate
the award of personal grants from the Walter Ahlström foundation, the
Foundation of Technical Progress (TES), the Emil Aaltonen foundation, the Tauno
Tönning foundation and the scholarship fund of Tyrnävä municipality.
I like to express my sincerest thanks to my supervisor Prof. Jouko Niinimäki
for his guidance and encouragement during the work. Special thanks go to Dr.
Esko Härkönen for advising and sharing his knowledge on mechanical pulping
with me. I would also like to thank Dr. Mats Ullmar, Mr. Petteri Vuorio and Mr.
Esa Viljakainen for many discussions, valuable comments and good advice during
I am most grateful to the reviewers of this thesis, Dr. Kari Luukko from
UPM-Kymmene and Prof. Bruno Lönnberg from Åbo Akademi. I would like to
thank Mr. Mark Jackson for proof reading the English language of this manuscript.
All my colleagues and friends I have worked with during these years in the
Fibre and Particle Engineering Laboratory are too acknowledged. Especially I
like to thank Mr. Esa Anttila, Mr. Sami Ojala, my dear sister Elisa Karjalainen, Mr.
Jarno Karvonen and Mr. Jani Österlund for their help in laboratory and
Finally, I like to thank those nearest and dearest to me who have encouraged
and supported me over the years.
Isä ja äiti, kiitos tuesta.
Sanna, Ansku ja Anna, kiitos rohkaisusta.
Mari, Helena, Pirjo ja Hanna, kiitos ystävyydestä.
Mikael, Joel, Enni, Joonas, Erika, Annika, Maija ja Iiro, kiitos kodista.
Oulu, September 2008
List of original publications
Illikainen M, Härkönen E, Ullmar M & Niinimäki J (2006) Distribution of power
dissipation in a TMP refiner plate gap. Paperi ja Puu 88(5): 293–297.
II Illikainen M, Härkönen E, Ullmar M & Niinimäki J (2007) Power consumption
distribution in a TMP refiner: comparison of the first and second stages. Tappi Journal
III Illikainen M, Härkönen E & Niinimäki J (2007) Power consumption and fibre
development in a TMP refiner plate gap: Comparison of unidirectional and standard
refiner segments. Proceedings of the 2007 International mechanical pulping
conference, Minneapolis, USA.
IV Illikainen M, Härkönen E, Ullmar M & Niinimäki J (2008) Disruptive shear stress in
spruce and pine TMP pulps. Paperi ja Puu 90(1): 47–52.
V Illikainen M & Niinimäki J (2007) Energy dissipation in a TMP refiner disc gap.
Proceedings of The 6th Biennal Johan Gullichsen Colloquium, Espoo, Finland: 49–57.
Theoretical calculations of Papers I, II and III were done and reported by the
present author. Experimental work presented in Papers III and IV were designed,
analysed and reported by the present author. Additional authors participated in the
designing and writing of papers by making valuable comments.
List of original publications
1.1 Background ..............................................................................................11
1.2 Problems to be studied ............................................................................ 12
1.3 The hypothesis ........................................................................................ 12
1.4 Initial assumptions .................................................................................. 12
1.5 Research environment............................................................................. 13
1.6 Outline of the thesis ................................................................................ 14
2 Refining mechanisms–present understanding
2.1 Flow phenomena inside the refiner ......................................................... 16
2.1.1 Pulp flow behaviour ..................................................................... 16
2.1.2 Pulp residence time in a disc gap.................................................. 18
2.2 Development of pulp and fibre properties in refining............................. 19
2.2.1 Separation of fibres....................................................................... 20
2.2.2 Development of pulp and fibre properties .................................... 21
2.2.3 Latency ......................................................................................... 22
2.3 Factors affecting fibre development........................................................ 23
2.3.1 Properties of wood raw material................................................... 23
2.3.2 Softening behaviour of wood ....................................................... 24
2.3.3 Specific energy consumption........................................................ 26
2.3.4 Intensity of refining ...................................................................... 27
2.4 Energy consumption of refining.............................................................. 28
2.4.1 Energy transfer and energy dissipation mechanisms .................... 29
2.4.2 Distribution of refining energy ..................................................... 31
3 Materials and methods
3.1 Equipment of shear and compression ..................................................... 33
3.2 Materials and methods used in experimental studies .............................. 35
3.2.1 Materials....................................................................................... 35
3.2.2 Compressibility of pulps............................................................... 36
3.2.3 Internal friction of pulps ............................................................... 37
3.2.4 Pulp pad refining–development of fibre properties in
fibre-to-fibre refining ................................................................... 39
3.3 Power consumption distribution in a refiner disc gap ............................. 41
3.3.1 Method.......................................................................................... 41
3.3.2 Mill trials ...................................................................................... 42
4.1 Compressing and shearing behaviour of pulps........................................ 45
4.2 Power consumption distribution in a refiner disc gap ............................. 54
5.1 Equipment of shear and compression and fibre-to-fibre refining............ 57
5.2 The method to calculate power consumption distribution ...................... 59
5.3 Mechanisms of refining – Why energy is dissipated?............................. 61
Thermomechanical pulp (TMP) is mainly used for mechanical printing papers
such as newsprints as well as uncoated and coated magazine papers. The
invention and development of thermomechanical pulping process in the 1970’s
rapidly changed the structure of the world’s pulp and paper industry. Previously,
printing papers were produced from a mixture of groundwood and kraft pulps.
Thermomechanical pulping rapidly displaced traditional mechanical pulping,
grinding, because similar paper properties could be achieved using fewer amounts
of expensive reinforced kraft pulp.
Today, thermomechanical pulping process is the most dominate mechanical
pulping process. The future of the process is, however, at stake. The continually
rising price of electricity and high electrical energy consumption of the process
have impaired the profitability of the process. TMP refining has been reported to
be the most energy intensive unit process in the pulp and paper industry (Sabourin
The heart of the thermomechanical pulping process is refining. Although
there are different refiner configurations used in the industry, depending on the
manufacture and process requirements, the main operational principle of all of
today’s refiners is the same: wood chips are fed between two parallel discs and at
least one of those is rotating. The patterned refiner segments transfer the
rotational energy into the pulp. Chips are broken down and due to compressive
and shearing forces a wood material is developed suitable for papermaking.
Refining process is very complex: conditions inside the refiner are very harsh,
flow phenomena inside the refiner are very complicated while two simultaneously
phenomena occur, separation and development of fibres. Studying the
mechanisms of refining is thus very challenging and until now it is not been
clearly understood what happens in a refiner disc gap: what is the desired
mechanism for efficient fibre development and how the energy is transformed
into pulp? However, it has been proposed in literature that refining energy could
be halved if the mechanisms of refining were known (Sundholm 1999a), because
theoretical energy requirement for creating new fibre surfaces is small. The
purpose of this work was to obtain new information on mechanisms of refining
and energy consumption of refining in the inner area of the refiner disc gap. This
was achieved by studying the power distribution in a refiner disc gap, as well as
fibre-to-fibre interaction in refining.
Problems to be studied
In order to understand mechanisms of refining it is extremely important to
identify and understand phenomena in the inner area of a disc gap. The residence
time of pulp in the inner area of disc gap is high and therefore it is likely to be one
important position that causes high energy consumption of refining. The role of
this area for energy consumption of refining is not well known. It is also not
known, if effective fibre development can occur in fibre-to-fibre contacts, which
is very likely to be the most important mechanism in the inner area of disc gap.
The aim of the thesis was to obtain new information about mechanisms of
refining and high energy consumption of the process. To solve the above
problems the following hypothesis were tested.
In the first stage refiner with standard refiner segments, the power
consumption is significant in the inner refining zone in a disc gap. (Papers I,
II and III).
The effective refining can be achieved in the fibre-to-fibre refining without
any impacts due to bar crossing (Paper V). The effectiveness of pulp pad
refining depends also on temperature and fines content of pulp (Paper IV).
The refining process and refiner type depends on mill and paper grade in question
and there are several different possibilities to produce thermomechanical pulp. In
this work the “standard” TMP line (Tienvieri et al. 1999) and the single disc
refiners in the first and the second stages have been used in calculations. The
proposals in literature about energy intensive inner refining zone inside the refiner
and also about importance of fibre-to-fibre mechanisms have also been connected
to the single disc refiners and so the results of experimental work can also be
applied mainly to the single disc TMP refiners.
In calculations some assumptions have been done. Mill trials and sampling
were done during normal operational conditions and it was assumed that the
measured values of consistency, temperature and the process variables well
describe the common situation inside the refiner. The steam was assumed to be
saturated in a disc gap. Water and fibres were assumed to be mixed and their
temperature was assumed to be equal.
Conditions inside the TMP refiner are very harsh and thus it is very
challenging to study refining mechanisms. To find a solution for this a piece of
equipment for shear and compression measurements was developed for studying
compression and shearing behaviour of pulp in a more controlled environment.
Compared to a situation in a refiner disc gap, some simplifications have been
made: in the experiments pulp was heated using steam, the compressive pressure
directly affected the pulp and the rotational speed of the equipment was very low.
Despite of these simplifications the study of compressing and shearing behaviour
of pulps certainly widens the present knowledge on mechanisms inside the real
The authors work about refining mechanisms consists of experimental and
theoretical parts. In the experimental section behaviour of pulp under
compressing and shearing forces was studied in the Fibre and Particle Laboratory
at the University of Oulu using apparatus for shear and compression experiments.
The raw material for experiments was produced at Metso Anjalankoski pilot plant.
Analysis of pulps was mainly preformed using laboratory facilities at the
University of Oulu and some at the Finnish Pulp and Paper Research Institute
The theoretical section was based on calculations and mill trials. Pulp
samples from the refiner disc gap were taken from the commercial SD-65 refiner
first and second stage positions at UPM-Kymmene Kajaani mill. Pulp samples
and process data collected during trials (used in Papers I and II ) were analysed at
the University of Oulu. The remaining test trials (Paper III) were performed at
UPM-Kymmene Kaipola mill in the first stage position using different refiner
segment types. The process data was collected and pulp analysis was done by
Outline of the thesis
The Chapter 1 of this thesis describes a short introduction to the study, the
problem to be investigated, the hypothesis while presenting the research
environment and an outline of the thesis. The Chapter 2 summarises previous
published studies on TMP refining mechanisms while the Chapter 3 presents the
materials and methods used in this study. The Chapter 4 summarises the main
results achieved during the work. The interpretation and discussion, as well as
proposals for further work are presented in Chapter 5. Finally, a conclusion is
presented in Chapter 6.
Refining mechanisms – present
In refining, moist wood chips and dilution water are fed into the first stage refiner.
The chips are broken down when they make contact with breaking bars of the
rotor. Due to rotation of the disc and pattern of the refiner segment surface,
rotational energy is transferred into the pulp by compression and frictional forces.
The outcome of the refiner is a number of different kinds of wood particles:
shives, coarse fibres, fibrillated and flexible fibres, as well as fines. The questions
arising here are how these different fractions have been generated and how much
energy is needed to create them.
The understanding of mechanisms of refining can be approached from
various perspectives. The following chapters present an understanding of the
refining phenomenon from three different approaches in order to get the best
possible view on what happens inside the refiner: 1) what are the main flow
phenomena’s occurring in a refiner disc gap based on visual and some theoretical
methods, 2) how fibre and pulp properties are developed during refining and what
factors are affecting fibre development and 3) energy consumption in a refiner
disc gap: how and where it is consumed. The remaining approaches are not
discussed here although the refining can be approached from the basis of
computational fluid dynamics (Huhtanen 2004, Härkönen et al. 1997) and
refining modelling (e.g. Berg & Karlström 2003, Corson 1973).
Flow phenomena, development of pulp properties, as well as energy
consumption of refining vary significantly depending on the radial position in a
refiner disc gap. Here, the disc gap is divided into three different zones, according
to Fig. 1. The most inner part of the refining is referred as breaking zone (Zone I),
the middle part of the refiner as inner refining zone (Zone II) and the narrow disc
gap as outer refining zone (Zone III).
Fig. 1. Refiner disc gap divided into sections.
Flow phenomena inside the refiner
Refining is a very complex process. Moist wood chips and dilution water are fed
into the centre of the discs. Water is needed to soften the wood (see Chapter 2.3.2)
and for protecting wood against burning. In a disc gap frictional forces cause the
evaporation of huge amounts of water into steam which is flowing outwards and
backwards. Flow direction and velocity of pulp depends on forces acting on pulp
(centrifugal force, frictional forces, and force due to flowing steam) that are of
different sizes in different positions inside the refiner. Pulp flow in the refiner has
been studied using different visual methods while a theoretical study about fibre
flows is presented (Chapter 2.1.1). The fibre residence time has also been studied
by theoretical methods and by experimental residence time measurements
2.1.1 Pulp flow behaviour
Pulp flow patterns have been studied using high-speed photography in both
atmospheric (Atack 1980, Atack et al. 1984, Stationwala 1992) and in pressurised
(Atack et al. 1989) refiners. The main findings of these studies are presented as
follows. Wood chips are shredded into coarse pulp immediately in the breaking
zone of the refiner (Zone I, Fig. 1) (Atack 1980, Atack et al. 1984). In the
breaking zone it has been shown that there is a considerable circulation of coarse
pulp and shives (Atack et al. 1984, Atack et al. 1989). Backflow was observed to
occur in the grooves of the stationary plate and forward along grooves of the
rotational plate (Atack et al. 1984). In the breaking zone of the disc gap (zone I,
Fig. 1) it has been reported that 70 to 80% of bar area is covered with pulp
(Stationwala 1992). In the inner refining zone (zone II, Fig. 1), the coverage of
pulp in the transition area between coarse and fine pattern was even higher. In the
outer refining zone (zone III, Fig. 1) the amount of coverage was lowest.
In the outer refining zone fibre flow was observed to be radiating outwards
(Atack 1980, Atack et al. 1984). Formed fibre flocs that moved tangentially, and
changing their shape and being disrupted were also observed in the process
(Stationwala 1992, Atack et al. 1989). The fibres and fibre flocs have also been
seen to be trapped on the bar edges where they are subjected to the refining action
(Atack 1980, Stationwala 1992, Atack et al. 1989).
Another research group (Alahautala et al. 1998) has also taken images inside
the refiner using a CCD-camera, light stroboscope and endoscope optics to view
the pulp more closely. The focus of their measurements was on pulp velocity, pulp
orientation, pulp coverage and the presence of fibre flocs. The results of their
studies showed that the back-flow of pulp in the breaking zone (zone I, Fig. 1)
was very slow. In the grooves within the inner refining zone (zone II, Fig. 1) there
was a pulp back-flow of 1 m/s while in the outer refining zone (zone III, Fig. 1) of
the refiner the radial velocity of pulp was as high as 30 m/s. The orientation of
fibres in the breaking zone was minor, but in the inner and outer refining zones
the fibres were tangentially orientated. The breaking zone was seen to be full of
pulp while in the inner refining zone average pulp coverage was below 60% and
below 10% in the outer refining zone.
Härkönen et al. (1997) have presented a theoretical model of refiner which
they have used to simulate the flow phenomena inside the refiner. Their computer
model consists of a steam phase and a combined fibre and water phase whilst
being based on mass, momentum and energy balance. The theoretical model has
proposed to give an idea of velocities of steam and water as well as volume
fraction of pulp in a disc gap. There is however some problems in the model in
that several variables have to be determined by trial and error in connection with
A calculated example using the model shows that volumetric density of pulp
(volume fraction of pulp) is highest in the inner refining zone (zone II, Fig. 1)
between the fine and coarse bar zones. This is analogous to visual studies of
Stationwala et al. (1992). The strongest turbulence in the pulp flow was observed
in the breaking zone of the refiner before it entered the plane section plate gap.
Qualitative diagrams of pulp and steam flows in a refiner, as well as volume
fraction of pulp in a refiner disc gap are presented in Figs. 2 and 3.
Fig. 2. Qualitative view of fibre and steam flows within the refiner (Härkönen et al.
Fig. 3. Qualitative view of volume fraction of pulp in a refiner disc gap (Härkönen et al.
2.1.2 Pulp residence time in a disc gap
In order to understand the mechanisms of refining the quantity of pulp in a refiner
disc gap as a function of disc radius should be known, in addition to how rapidly
the pulp is flowing at a given point. The residence time of pulp in different parts
of the refiner disc gap is essentially important to know because it indicates how
much pulp is available for refining action.
Miles & May (1990) proposed a theoretical model to calculate pulp residence
time in a refiner disc gap. They derived an equation for velocity of pulp in a
refiner disc gap based on forces affecting pulp. Their model has, however, some
assumptions which the model correctness has been questioned by e.g. Härkönen
et al. (1997) and Sabourin et al. (2001). In addition there are many coefficients
that are difficult to measure. The residence time model presented by Miles and
May has later been shown (Ouellet et al. 1996, Härkönen et al. 1999, Murton &
Duffy 2005) not to predict the measured residence time of pulp in a refiner disc
The quantity of pulp has been observed to be at its highest in the inner parts
of the refiner (Stationwala 1992) (zones I and II, Fig. 1) which was explained by
recirculation of pulp in the inner parts of the refiner and restriction of the radial
pulp flow caused by the fine plate pattern. This indicates the pulp’ residence time
being highest in the inner parts of the refiner. The residence time of pulp inside
the refiner has later been measured using a radioactive tracer in a commercial
TMP refiner (Härkönen et al. 1999, Härkönen et al. 2003, Murton & Duffy 2005).
In these studies the pulp velocity in the outer parts of the refiner was shown to be
high compared to the inner parts. Härkönen et al. (1999) reported that residence
time of pulp in the areas of breaker bar and centre segments was between 2.5 to 7
s (zones I and II, Fig. 1) while at the outer part of the plate gap it was
approximately 0.5 s (zone III, Fig. 1). Murton & Duffy (2005) measured similar
residence times: 0.5 s in the outer refining zone (zone III, Fig. 1) and 3–15 s
within the refiner system (ribbon screw to blow valve). A number of reports have
suggested the possibility of changing the pulp’s residence time by adujsting the
segment geometry (Härkönen et al. 1999), by controlling the consistency and the
pressure differential over the disc gap (Härkönen et al. 2003)
Development of pulp and fibre properties in refining
The main purpose of thermomechanical pulping process, as well as all other
pulping processes, is to separate the fibres from the wood and to make them
suitable for papermaking. Sundholm (1999b) has suggested that in an ideal
mechanical pulping process the following phenomena should happen: fibres must
be separated from wood matrix, fibre length must be retained, fibres must be
delaminated, abundant fines must be generated by peeling off outer layers of the
middle lamella, primary and secondary layers of fibre wall, and finally the surface
of the remaining secondary wall must be fibrillated.
In the thermomechanical pulp refiner the separation of fibres and
development of fibre properties occurs simultaneously but are considered
separately in the following chapters. The latency of pulp is also discussed in a
separate chapter due to its importance in the author’s experimental work on the
development of pulp in fibre-to-fibre refining (Paper V).
2.2.1 Separation of fibres
It has been shown that wood chips are broken down immediately into coarse pulp
in the breaking zone of the refiner (Atack et al. 1984). Further separation occurs
when coarse pulp is affected by frictional forces in refiner disc gap so the shive
content of pulp decreases as a function of radius of the disc gap (Atack et al. 1984,
Härkönen et al. 2003).
The exact mechanism of fibre separation is not understood and this is
highlighted by the very low amount of publications available about the initial
breakdown of chips or by the separation of fibres. This is probably due to the
complicated and simultaneous process of the fibre separation and development
(Fundamental studies on disintegration of wood under different loading modes
has been preformed by Koran 1968, Law & Koran 1981, Koran & Salmén 1985).
In refining, separation of fibres occurs in the weakest part of the wood which is
dependent on moisture, temperature and frequency (see Chapter 3.2.2), but pretreatment methods (chemical, mechanical, biological) could also be possible
factors. In refining, fibres are separated from each other through the primary wall
– secondary wall interface or in the secondary wall (Gustafsson et al. 2003).
An initial separation step has been shown to be important factor for final pulp
quality. Heikkurinen et al. (1993) have analysed pulps produced from using small
amounts of energy (SEC 0.5 MWh/t) and they observed that conditions
(temperature and rotational speed) in the initial breakdown process had a major
effect on size and quality of undefiberised particles. Falk et al. (1987) tested
single and double-disc refiners in the first and second stages and found that pulp
properties were determined by the conditions in the first stage position. Rudie &
Sabourin (2003) found that there are differences in wood breakdown rates
between different types of wood (plantation and forest-grown loblolly pine wood).
Different types of fibres have also been shown to behave differently in the
defibration step. Murton et al. (2006) have concluded that in the fibre separation
stage, thick-walled latewood type fibres are liberated first and more easily than
large diameter, thin-walled earlywood fibres. They have also showed that with
low specific energy consumption levels the shive content decreases rapidly while
later in refining shive content reduces gradually. They proposed that individual
liberated fibres prevent the refining energy to be applied directly to shives.
2.2.2 Development of pulp and fibre properties
Careful studies have been made on the development of pulp and fibre properties
during thermomechanical pulp refining. Development of fibre and pulp properties
has been studied in the refiner disc gap, as a function of disc radius, or as a
function of specific energy input (the term “more refined pulp” means pulp with
higher specific energy input).
Development and understanding of pulp properties in a refiner disc gap has
been studied by taking many samples from within this area. First attempts were
preformed in the 1980’s using an atmospheric 1.9 MW refiner (Atack et al. 1984)
and more recently Härkönen et al. studied pulp properties using the 10 MW
pressurized SD-65-refiner (Härkönen et al. 2003). Their studies showed that pulp
properties are developed in the breaking zone (I, Fig. 1), in the inner refining zone
(II, Fig. 1), and in the outer refining zone (III, Fig. 1). According to these studies
in the inner parts of the refiner there is no fibre cutting however defibration of
shives and chips does occur. In the parallel disc gap the fibre length is reduced
and fine material is created through the harsh refining.
The increasing specific energy consumption of refining changes pulp and
fibre properties. Pulp fractional composition changes when energy input is
increased: the amount of long fibre fraction decreases while the amount of fine
material increases (Law 2000), shive content decreases (Karnis 1994), the average
fibre length, as well as pulp freeness decreases (Corson 1989, Tienvieri et al.
Properties of different fibre fractions also change. The coarseness of long
fibre fraction decreases (Stationwala et al. 1996, Law 2000, Karnis 1994) and
fibres become more flexible (Corson 1989, Karnis 1994). Fibres are peeled,
fibrillated (Moss & Heikkurinen 2003), collapsed and their cell wall thickness
decreases (Corson & Ekstam 1994, Kure & Dahlqvist 1998, Jang et al. 1996, Jang
et al. 2001). Internal properties of fibres also change, allowing the fibre cell wall
to swell, and thus swelling increases when increasing the amount of refining
(Moss & Heikkurinen 2003). In addition, properties of fine material changes too
so that the specific surface area of fines increases when increasing the SEC
(Stationwala et al. 1996). It has also been shown that by peeling the outer layers
of fibres fines are generated (Heikkurinen & Hattula 1993, Luukko & Paulapuro
1999, Luukko & Nurminen 1999). In early stages of refining fines are generated
from outer layers resulting in fines having good light scattering properties. Later
in the process, properties of fines correspond to properties of inner fibre wall (S2
layer). In addition to property alternations presented above, refining can cause
different fibre damage like cracks in the fibre wall (Gregersen & Holmstad 2004).
The development of fibres has been shown to occur by the unravelling and
peeling of outer layers of fibres (Forcags 1963, Karnis 1994, Stationwala et al.
1996, Heikkurinen & Hattula 1993) while, cutting of fibres is seen in refining.
(Law 2005). A good example of a figure showing the mechanisms through which
the development of refining occurs has been presented by Karnis (1994) (Fig. 4).
Initial breakdown of chips to shives occurs by disintegration of particles, fibre
development by disintegration and peeling mechanisms, the latter (peeling) being
the dominating mechanism in refining.
Fig. 4. Separation and development of fibres in refining (Karnis 1994).
A special phenomenon occurring in thermomechanical pulp refining is the
formation of latency which generates curled and twisted fibres. The most visible
effect is the decrease of freeness during hot-disintegration (SCAN-M 10:77).
Beath et al. (1966) were the first to show that freeness of disc refined pulp
decreased significantly during hot-disintegration. They understood that the refiner
produces pulp with lower freeness but somehow freeness is increased in the
process. They found part of pulp properties being effectively latent and thus it is
why they referred to this property as latency. Beath et al. also proposed that
latency is due to twisting and bending of pulp fibres at high temperature
conditions inside the refiner. When pulp cools, fibres remain in their curled form.
Jones (1966) published microscopic pictures on pulps before and after hotdisintegration. He showed that curled kinked and twisted mechanical pulp fibres
can be straightened in hot-disintegration.
Removal of latency is extremely important in order to obtain reliable results
in pulp testing. Hot-disintegration is the most common way to remove latency of
mechanical pulps. Besides that, beating has been reported to be an additional
suitable method (Beath et al. 1966). The removal of latency not only decreases
freeness but increases a burst factor (Beath et al. 1966). Removal of latency has
also been reported to decrease shive content and increase tensile strength (Mohlin
Factors affecting fibre development
In refining, there are many factors affecting fibre separation and refining result
which are discussed in this chapter. Properties of wood raw material give a
framework for final pulp properties. Temperature and moisture in refining are key
factors affecting fibre separation and development due to softening and
viscoelastic behaviour of wood material. Besides to these, refining results are
strongly dependent on two important factors: how much and how the energy is
transferred into pulp.
2.3.1 Properties of wood raw material
Wood species is the most important raw material parameter in thermomechanical
pulp refining. Different wood species result in different final pulp properties and
also differences in specific energy consumption. (e.g. Härkönen et al. 1989,
Hatton & Johal 1996, Reme & Helle 2001). The most favourable raw material for
thermomechanical pulping is spruce, especially Norway spruce that has
favourable fibre properties, low extractive content and high initial brightness of
wood (Varhimo & Tuovinen 1999). Most pine species consume much more
energy compared to spruce species, and pulp properties are not as good.
Wood is an anisotropic material; its properties vary a lot between different
species but also within a certain wood species themselves. The refining result is
thus dependent on many other parameters, e.g. moisture content of wood, wood
basic density and seasonal variations (earlywood and latewood fibres) (e.g.
Fuglem et al. 2003, Reme et al. 1999, Mohlin, U.B. 1997, Tyrväinen 1997,
Murton & Corson 1992, Corson 1991).
2.3.2 Softening behaviour of wood
The fracture of wood (separation and development of fibres) in mechanical stress
occurs in the weakest parts of the wood structure which is defined by the moisture,
temperature and strain rate. Wood consists mainly of three viscoelastic polymer
materials, cellulose, hemicellulose and lignin, and their softening behaviour
defines how wood material is broken down. The proportion of these polymers
varies in fibre cell wall structure, lignin being the main component in the middle
lamella (Fig. 5) (Panshin & De Zeeuw 1964).
In refining conditions wood is saturated with water. In water-saturated
conditions at 20 ºC cellulose and hemicellulose are already softened (Back &
Salmén 1982). Thus lignin softening is a critical factor from the fibre separation
point of view. Lignin is a viscoelastic material and at low temperatures (below the
glass transition) it is stiff and glassy. In transition region the stiffness decreases
and at high temperatures lignin behaves like a rubbery material (Irvine 1984). In
refining, if lignin is too stiff the fracture occurs throughout more soften fibre wall
and the refining results in broken fibres and high fines content (Salmén et al.
1999). If lignin is too soft refining produces pulp covered in a layer of lignin and
it is impossible to refiner any further (Atack 1972).
Softening of lignin is dependent on strain rate or strain frequency. At low
strain rates, lignin readily softens at about 70 ºC (Salmén et al. 1984) while the
softening temperature increases at higher strain rates. It has been reported that
refining conditions between 120–135 ºC is needed for the lignin to become soft
(Atack 1972) but it has also be reported that the softening temperature range is
much wider, from 100 to 170 ºC. The optimal refining occurs in lignin softening
transition zone where lignin is not glass-like anymore but not yet become a total
rubbery material. In optimal thermomechanical pulp refining, fibres are separated
from each other through a primary wall or through a primary wall – secondary
wall interface (Gustafsson et al. 2003), as illustrated schematically in Fig. 6.
Fig. 5. Distribution of the principal chemical constitutes within the various layers of
the cell wall in conifers (Panshin & de Zeeuw 1964).
Fig. 6. Breakdown of the wood matrix as a function of refining temperature. RMP 20–
95ºC, TMP 110–150ºC, MDF 170–190ºC (Franzén 1986).
2.3.3 Specific energy consumption
How much energy is put into pulp is defined as specific energy consumption
(SEC). In refining, it is determined as energy per unit mass of pulp, generally
calculated from refining power and production rate of pulp (Eq. 1). The unit of
SEC normally used is MWh/t. The more energy is put into pulp the more refined
the pulp is and greater the changes in fibre and pulp properties are (see Chapter
3.2.2) However, with the same SEC totally different kinds of pulps can be
produced while the other important factor in refining is how the energy is put into
pulp, commonly referred as a “intensity of refining”
P is the power (MW),
m is the mass flow rate (t/h).
2.3.4 Intensity of refining
The concept of refining intensity describes how rapidly energy is put into pulp. It
has a theoretical background, presented in the following paragraph. The
theoretical concept of refining intensity has, however, some limits and it is thus
more appropriate to use refining intensity as a qualitative term to describe if the
refining is harsh or gentle.
The concept of refining intensity was adopted in thermomechanical pulp
refining by Miles and May in the 1990’s (Miles & May 1990, Miles 1990). They
proposed that the important factor affecting refining phenomenon is a number of
impacts the pulp is receiving during refining. This is caused by refiner bar
crossing and how much energy is transferred into pulp during a particular bar
crossing. They derived an equation for radial velocity of pulp in a refiner disc gap,
based on forces acting on pulp. This equation was then used to calculate pulp
residence time in a refiner disc gap, and consequently for calculating the number
of impacts pulp is receiving in a disc gap. Refining intensity was then determined
according to Eq. 2 and it was proposed to be an important factor affecting the
e is the refining intensity (MWh),
E is the refining energy (MWh),
n is the number of impacts.
The concept of refining intensity was rapidly adopted into research of
thermomechanical pulp refining. The effect of theoretically calculated refining
intensity on pulp properties and energy consumption has been studied in several
papers. The effect of refiner speed (e. g. Miles & Karnis 1991, Miles & Omholt
2003), as well as effects of plate pattern (Murton 1998, Murton et al. 2002),
consistency (Alami et al. 1997) and production rate (Murton & Corson 1997) on
refining intensity, energy consumption and pulp properties has been studied.
According to studies presented above, the increasing refining intensity
decreases energy consumption of refining and the result, in general, is pulp with
lower amounts of long fibres, shorter average fibre length, more fines, lower
strength properties and better optical properties. In some publications it has been
shown that a reduction of energy consumption can be achieved with high intensity
refining without any significant changes in pulp properties (Kure et al. 1999
Sabourin et al. 2001, Sabourin et al. 2003).
The theoretical determination of refining intensity is not exactly correct. In
the model of Miles and May there are several factors that are difficult to
determine. Pulp residence time measurements (presented in Chapter 1.1.2) have
also shown that the theoretical residence time of pulp (determining the refining
intensity) do not correspond to those determined experimentally. The calculated
refining intensity has also been shown to have a limit when directional plate
patterns were used (Kure et al. 1999). This indicates that pulp residence time,
more than a number of refiner bar impacts, affects pulp quality. The comparison
of high intensity (RTS: high speed, temperature, low residence time) and normal
TMP processes, Sabourin et al. (2001) concluded that theoretically determined
refining intensity does not correspond to the quality of pulp: in their experiments
similar pulp properties were obtained using low and high refining intensities.
Because there are limitations in theoretical refining intensity, the concept
should be rather used as a qualitative term: to describe if the refining is harsh or
gentle. High intensity refining can be understood as a harsh refining, or more
rapid refining (with lower residence time) compared to gentle refining. Harsh
refining reduces fibre length, creates fine material and reduces long fibre fraction
whereas gentle refining defibrates fibres without fibre cutting.
Energy consumption of refining
TMP refiner has been shown to work ineffectively from the fibre development
point of view (Law 2004) and the energy saving potential of the process has been
reported to be higher than in any other unit process in pulp and paper industry
(Münster et al. 2003). Uhmeier and Salmén (1996) have surveyed that the
theoretical energy needed to produce new fibre surfaces has been estimated from
less than 1% to 30% of the total energy consumption of pulp. They reviewed the
energy needed to develop fibres suitable for papermaking and proposed it to be
much higher compared to the separation energy. They also proposed that a large
amount of energy is wasted due to other losses which do not contribute to
transforming the wood chips into a useful pulp (Uhmeier & Salmén 1996).
Reasons for high energy consumption of TMP process are not known
although some proposals have been put forward. The present understanding and
opinions about energy transfer and energy dissipation mechanisms are presented
in Chapter 2.4.1. Chapter 2.4.2 presents studies on energy distribution in a refiner
2.4.1 Energy transfer and energy dissipation mechanisms
The purpose of refining segment surface is to transfer the rotational energy of the
refiner disc into the pulp. The contacts between segment surface and pulp are thus
extremely important. For energy transfer and fibre development the other
interaction, fibre-to-fibre interaction, is important as well.
A generally accepted theory about refining mechanism is that fibres are
subjected to repeated compressions and decompressions caused by bar crossing
and development of fibres occurs by process of wood fatigue. Salmén et al. (1985)
has performed wood fatigue tests observing similar structural changes of fibres in
wood matrix and those occuring for long fibre fraction in refining. Since then the
refining has often been described as a fatigue phenomenon (e.g., Kurdin 1998,
Salmén et al. 1997, Uhmeier & Salmén 1996, Wild et al. 1999). The number of
refiner bar impacts is believed to be the critical factor affecting the development
of fibres and energy consumed in refining (Miles & May 1990) while the concept
of refining intensity (energy per impact) has been widely used to describe the
refining phenomenon (See Chapter 3.3.4). Uhmeier and Salmén (1996) have
proposed that a reason for high energy consumption of mechanical pulping could
be the large number of large radial compressions. They studied repeated large
radial compressions of heated spruce to determine the most efficient way to
achieve the desired fibre changes. They suggested that a small number of large
compressions efficiently increased the collapsibility of fibres while a large
number of large compressions were observed to be an inefficient way to cause
The purpose of refiner segment surface can be seen, however, also as a
frictional surface that transfer the energy inside the compressed pulp pad where
refining occurs by shearing of pulp. Härkönen et al. (2003) proposed that shear
stress mechanisms occurrs in a disc gap. In this model the segment surface works
as a frictional surface transferring refining energy (or not) inside the pulp. Thus
energy transfer occurs by frictional forces inside the pulp or in the surface
between pulp and segment surface. Fig. 7 presents different possible shear stress
mechanisms. Continuous shear deformation in a fibre bed (Case 1), or at the
segment surface (Case 2) creates power consuming mechanisms in a disc gap.
Case 3 has been proposed to describe the situation in reality.
Tangential speed of rotor
in fibre bed,
of fibre against
in fibre bed, hitting
of fibre against
in plate gap,
of cases 1 and 2
Fig. 7. Shear stress mechanisms in a refiner disc gap (Härkönen 2003).
According to Kurdin (1986) a fibre mat is needed in the refiner disc gap. He
reported that optimum refining is accomplished in fibre-to-fibre action as opposed
to plate-to fibre action. Härkönen et al. (2003) proposed that in the inner parts of
the refiner (Zones I and II in Fig. 1), the power is mainly consumed in fibre-tofibre refining, by shearing of compressed pulp (Case 1 in Fig. 7) while in the
outer parts contacts between segment surface and pulp became more important.
Murton & Duffy (2005) have also proposed the mechanism in which inter-fibre
contacts are important. They proposed that a increased in refining energy input
requires a reduction in the plate gap causing a “throttling effect” which causes
pulp volume to decrease (pulp network density to increase) and increase interfibre contacts and friction between them. It has also been reported that inter-fibre
contacts are harmful in refining. Law (2000) proposed that formation of fibre
aggregates in the refiner disc gap reduces the efficiency of refining: aggregation
shells could work as a shield protecting the inner component from the shearing
action of refiner bars.
Although the fibre-to-fibre interaction is believed to be important for energy
consumption during refining, there are no publications available that clearly show
what role of this mechanism is for energy consumption and the result of the
refining process. In this work shearing of compressed pulp pad was studied and
the development of pulp properties in pulp pad refining was clarified.
2.4.2 Distribution of refining energy
Due to the very intricate nature of the refining process and complex flows inside
the refiner (Chapter 2.1 presents main fibre flows) it is essential to know in which
part of the disc gap energy is consumed in order to understand the high energy
consumption of the process. Miles and May (1990) assumed that energy of
refining is consumed mainly in the refining zone, energy consumption in the inner
parts being minor. Härkönen et al. (2001) have reported, based on their studies,
that energy consumption before the temperature maximum inside the refiner is as
large as the energy consumption in the outer parts of the refiner. A qualitative
picture about energy distribution in a refiner disc gap is presented in Fig. 8.
Measurements of shear forces in a refiner disc gap, in a laboratory refiner by
Gradin et al. (1997), and in a commercial refiner by Backlund et al. (2003), have
proposed energy consumption to increase along the disc radius. Backlund et al.
(2003) noticed more scattered forces in the inner parts of the refiner and proposed
that refining mechanism is different there compared to outer parts of the refiner.
Fig. 8. A qualitative picture about power consumption distribution presented by
Härkönen et al. (2001).
As presented above there are different, even contradictory estimations about the
proposed energy distribution. In this work a new method to calculate the energy
consumption distribution is offered and energy consumption distribution in a
refiner disc gap is calculated.
Materials and methods
The authors study about refining mechanisms has been experimental and
theoretical. In experimental work the internal frictional properties of compressed
pulps were studied and development of pulp properties in fibre-to-fibre refining,
or pulp pad refining, was clarified. In theoretical work the methods to calculate
power consumption distribution in the refiner disc was presented. In this chapter
equipment of shear and compression used in experimental work is presented (3.1),
as well as materials and methods (3.2) used in experimental work. Theoretical
methods are also introduced as well as mill trials in which these calculations are
based on (3.3).
Equipment of shear and compression
Equipment of shear and compression (ESCO) was developed during this work in
the Fibre and Particle Engineering Laboratory. The ESCO was used to study
compressibility of TMP pulps, internal frictional properties of compressed pulps,
friction between pulp and different surfaces, as well as development of pulps
under compressing and shearing forces (pulp pad refining). ESCO is presented in
Fig. 9. Equipment of shear and compression (ESCO).
Main parts of the equipment consist of a cylindrical vessel of 152 mm diameter, a
pneumatic press and a rotating bottom of the vessel. The device is connected to a
steam line and high pressure steam is used to heat up the pulp. A computer is used
for operating the equipment and for collecting and storing all the measured data:
temperatures inside the cylinder, piston position, compressive force and torque.
The main operational principles of the equipment are presented in Fig. 10 into
five operational principles. 1. Pulp is placed into the cylinder, and the base and
top of the cylinder are bolted in position. 2. The pulp is heated using steam
flowing through the pulp pad and is also present through the experiment in order
to keep temperature at its target value. The temperature of the steam is controlled
using a pressure-reducing valve. Temperatures up to 180 degrees are possible.
There is also a valve for condensed water on the lower side of the vessel. 3. The
pulp is compressed using the piston and the pneumatic press, during which the
upper and lower sides of the cylinder are connected for pressure balancing. 4. In
the shearing test the pulp is first compressed and then the bottom plate is rotated
so that the compressed pulp pad is disrupted. 5. In refining tests the pulp is first
compressed and then the bottom is rotated continuously for a desired time. The
motor is connected to the axis of the bottom plate and is used for rotating the
bottom. The torque required for disruption is measured. Rotational speeds from 5
to 84 rpm can be used.
Fig. 10. Operational principles of the ESCO.
The equipment can be assembled using different surfaces on the piston and
bottom plates. Smooth surfaces are used to study the compressibility of pulps,
while the bottom of the cylinder and the piston are fitted with needles for studying
shearing properties. Bar type segments can be used when studying frictional
properties between segment surface and pulp. Needle plates and bar type
segments are presented in Fig. 11.
Fig. 11. Needle plates (left) and bar type segment plates (right).
Materials and methods used in experimental studies
Thermomechanical pulps with different freeness values produced from Norway
spruce (Picea abies) chips and from both sawmill and thinned Scots pine (Pinus
sylvestris) chips were refined for experiments using a 20-inch pilot refiner in
Metso Anjalankoski pilot plant. Freeness levels and specific energy consumptions
are presented in Table 1. Table 1 also summarises materials used in different
experiments (compressibility, internal friction of pulp, fibre-to-segment friction,
pulp pad refining) and papers in which details of different experiments are
Table 2. A list of thermomechanical pulps used in experimental work. Spruce is
Norway spruce (Picea abies), pine is Scots pine (Pinus sylvestris).
3.2.2 Compressibility of pulps
In compressibility tests 100 g of absolute dry pulp was loaded into the cylinder at
a consistency level of about 30% and heated with steam at temperatures of 120,
150 and 170°C for 15 minutes or soaked with water and left at room temperature
of 15°C. Pulps were compressed using compressive pressures of 1, 2, 4, 6 and 8
bars. It has been reported that the average fibre force (force needed to balance the
sum of axial thrust and steam pressure round the rotor). in a refiner disc gap is
about 0.5 bars (Härkönen 1993) and because fibre pressure is certainly not
distributed evenly in a disc gap, higher pressures were used in this study.
Eq. 3 incorporates the definition of the volume fraction of pulp. The density
of the fibre cell wall is used to calculate the volume of solid fibre material, which
is then compared with the total volume of pulp
ρ fibre wall
ε is the volume fraction of pulp,
V is the volume [m3],
m is the mass of pulp [kg] and
ρ is the density of fibre wall [kg/m3].
3.2.3 Internal friction of pulps
Internal frictional properties of compressed pulp pad were studied as a disruptive
shear stress. The disruptive shear stress was measured as a function of
compressive pressure or volume fraction of pulp at different temperatures. Two
test points were also measured using pulp from which the fine fraction (P200) had
been removed by Bauer McNett fractionation. The amount of pulp in the
experiments was from 75 to 130 g of absolute dry pulp, depending on the pressure
used. The amount of pulp was chosen so that the distance between the needles
was about 5 mm in all experiments, as calculated using the results of the
compressibility tests. A heating time of 15 minutes was also used in the shearing
experiments. After heating, the pulp was compressed at a certain pressure, the
compressed pulp pad was then disrupted and the torque needed for this was
measured. The rotational speed of the bottom of the cylinder was 8.4 min-1.
The compressed pulp pad was disrupted by rotating the bottom of the vessel.
It was assumed that pulp pad behaves like an elastic material before disruption;
therefore Hooke’s law (Beer et al. 1992) was used to describe its behaviour. The
compressed pulp pad was considered as a rigid circular shaft. The shear stress in
the rigid shaft varies linearly with the distance from the axis of the shaft (Fig. 12).
The disruption starts when the highest value of shear stress is achieved. The
maximum shear stress in a pulp pad as a function of torque is presented in Eq. 4
(Beer et al. 1992) and is referred here as disruptive shear stress. After disruption
shearing surfaces slide against each other and shear stress is then more likely to
be uniformly distributed.
Fig. 12. Shear stress in the rigid shaft.
τ max =
Tmax ⋅ rc
is the disruptive shear stress [N/m2],
Tmax is the torque [Nm],
r is the radius of the cylinder [m] amd 4
π ⋅ rc
J is the polar moment of inertia [m4],
Friction between pulp and different surfaces was studied using various types of
surfaces on the piston and bottom plates: smooth surface, needle surface and bar
type segment surface, according to Fig. 13. Pine TMP pulp with freeness value of
720 ml was compressed between the piston and bottom plates, the torque needed
to disrupt the compressed pulp pad was measured and the disruptive shear stress
was calculated using Eq. 4. Experiments were done using compressive pressures
of 1, 2, 3 and 4 bars at 120 ºC.
Bar type segments
Fig. 13. Different surfaces used in experiments.
3.2.4 Pulp pad refining – development of fibre properties in fibre-tofibre refining
For the pulp pad refining experiments the ESCO was assembled using needle
plates (Fig. 11) on the bottom and piston plates. 220 g of absolute dry pulp (pilot
mill refined the first stage pulp with freeness value of 750 ml) was loaded and
then the bottom and top of the cylinder were bolted. The pulp was heated for 15
minutes using steam at a temperature of 150ºC, after which the pulp was
compressed using a certain pressure of the pneumatic cylinder. After compression
the position of the piston plate was locked and the motor connected to the bottom
of the cylinder was started. The bottom plate was rotated continuously for
different lengths of time. Steam was present during the experiment and it worked
not only as a heater but also as a cooler when steam was generated due to friction
during the refining experiment. The torque, distance between needles and
compressive pressure were measured. After the experiment the cylinder was
opened, the piston was elevated and pulp was collected from between the tops of
the needles for laboratory analysis (Fig. 14). Coarse, non-refined pulp was
observed in the needles, while the pulp between the tops of the needles was
visibly refined. An additional curl-experiment was performed using pulp with a
freeness value of 120 ml to clarify the effect of shearing treatment on the freeness
value of the pulp.
Fig. 14. Pulp was collected from between the tops of the needles after the experiments.
Pulp collected from between the tops of the needles after the experiment was
weighed, its dry content measured and the following analyses carried out after hot
disintegration (SCAN-M 10:77): Bauer McNett (BMN) classification (SCAN-M
6:69), Canadian standard freeness, CSF (SCAN-C21:65) and fibre length and curl
index using the FiberLab-analyser. Scanning electron microscopic (SEM) pictures
were taken of various BMN fractions, and the so-called “band” microscopic
method (Heinemann 2006) was used to study the peeling of the outer layers of the
fibres. Laboratory sheets were made from a pulp mixture (pulps from three
different refining experiments were mixed together) and tensile strength analysis
(SCAN-P38) was done.
Specific energy consumption
Specific energy consumption (SEC) during the experiment was estimated using
P is the power [W],
t is the time [s],
T is the torque [Nm] and
ω is the angular velocity [1/s],
Pt T ωt T 2π nt
n is the rotational speed [1/s],
m is the mass of the pulp between the needles [kg].
Power consumption distribution in a refiner disc gap
The calculation of the power consumption distribution is based on mass and
energy balances, and also on mill trials during which consistency and temperature
profiles inside the refiner are determined, as well as the production rate of the
refiner and the actual motor power. Refiner disc gap is divided into sections
according to sampling points (Fig. 15) and power consumption which is
calculated separately in different sections. The mass balance (Eq. 6, 7, and 8) is
used to calculate the mass flow through the sections. Enthalpy values of water,
fibres and steam, mass flows and temperature profile are used to calculate the
power consumption in different sections by energy balance (Eq. 9). Steam, fibres
and water are assumed to have equal temperatures while steam is also assumed to
be saturated in a disc gap. Boundary of the calculation is either the stagnation
point where steam velocity is assumed to be zero or the feed of the refiner where
amount of blow back steam is calculated. The consistency of pulp in the feed of
the refiner was not determined in connection to mill trials. It was either calculated
Fig. 15. Refiner disc gap divided into sections according to sampling points.
mi = mi +1 ,
(100 − ci )
⋅ mi ,
Si +1 = Si + Fi − i +1 ,
mi is the pulp flow at point i [kg/s],
Wi is the water flow at point i [kg/s],
C i is the consistency [%],
Si is the steam flow [kg/s],
Fi − i +1 is the mass flow between steam and water phases [kg/s].
Pi - i +1 = H m(i +1) ⋅ m (i +1) − H mi ⋅ mi
+ H W(i +1) ⋅ W(i +1) − H Wi ⋅ Wi
+ HS(i +1) ⋅ S(i +1) − HSi ⋅ Si ,
Pi - i +1 is the power consumption between points i and i + 1 [kW],
H mi is the enthalpy of pulp at point i [kJ/kg],
H Wi is the enthalpy of water at point i [kJ/kg],
HSi is the enthalpy of steam at point i [kJ/kg].
The total power consumption ( Ptot ) is the sum of partial power consumption of
sections, as presented in Eq. 10.
Ptot = ∑ Pi - i +1 .
3.3.2 Mill trials
Calculations were based on three series of mill trials presented below:
Papers I and II: Mill trials during 2005 at UPM-Kymmene Kajaani mill. The
SD-65 refiner, at the first stage position was used in trials. The sampling was
repeated five times to eliminate process variations. Temperature profile was
not measured. Production rate was 3.0 kg/h, actual motor power 10.5 MW.
Paper II: Mill trials during 2006 at UPM-Kymmene Kajaani mill. The SD-65
refiner, at the second stage position was used in trial. The sampling was
repeated five times to eliminate process variations. Temperature profile was
not measured. Production rate was 3.0 kg/h, actual motor power 6.3 MW.
Paper III: Mill trials between 1994–1995 at UPM-Kymmene Kaipola mill.
The SD-65 type single disc refiner, at the first stage position was used in the
trials: in one test trial standard refiner segments were used whereas in the
other trial the low energy (LE) segments were used (Fig. 16). Temperature
was measured using PT100-sensors. Temperature profiles are presented in Fig.
17. For LE-segments the motor power was 9.2 MW, production rate was 2.3
kg/s. For standard segments the power was 9.7 MW, production rate was 2.43
kg/s. The SEC was the same in both trials. Sampling was preformed only
once and pulp properties were measured.
Fig. 16. Standard (left) and LE (right) segments.
Fig. 17. Temperature profile in a disc gap of the first stage refiner equipped with
standard and LE-segments.
Compressing and shearing behaviour of pulps
Pulps were more compressible at high temperatures (120–170 °C) compared to
room temperature (Fig. 18) due to softening of wood. No differences were
observed between higher temperatures indicating that fibres were already totally
softened at 120 °C. Internal friction of the compressed pulp pad was observed to
be affected by temperature (Fig. 19). The lower the temperature became, higher
the disruptive shear stress was. Differences between coarse (720 ml, 540 ml) and
fine pulps (160 ml) were also observed: the internal friction was lower in pulps
with a lower freeness value which indicated that fines work as a kind of lubricant
in shearing. The highest disruptive shear stress was achieved when the fine
material was removed by Bauer McNett fractionation, as can be seen in Fig. 20.
Fig. 18. Compressibility of Scots pine (Pinus sylvestris) thinning TMP pulp (CSF 720
ml) at different temperatures.
Disruptive shear stress [N/m2]
Fig. 19. Disruptive shear stress as a function of the volume fraction of pulp at different
temperatures. Norway spruce (Picea abies), CSF 300 ml.
Disruptive shear stress [N/m2]
sawmill pine 156 ml
sawmill pine 340 ml
sawmill pine 480 ml
sawmill pine 723 ml
sawmill pine - fines removed
Fig. 20. Disruptive shear stress as a function of volume fraction of pulp. The effect of
freeness and fines. Scots pine (Pinus sylvestris).
The effect of segment surface on frictional properties, e.g., disruptive shear stress
of compressed pulps, was measured using different segment surfaces in the ESCO.
The total disruption was achieved only if needle plates were used in equipment, as
can be seen in Fig. 21. Using a smooth surface the lowest disruptive shear stress
was measured. Using bar type surfaces the disruptive shear stress was between
values obtained for smooth and needle surfaces.
Disruptice shear stress [N/m2]
Compressive pressure [bar]
Fig. 21. Disruptive shear stress as a function of compressive pressure. Different
segment surfaces were used.
In pulp pad refining experiments, the development of pulp and fibre properties as
a function of specific energy consumption was studied. The main results are
collected in Figs. 22–27. The pulp properties were compared to mill refined pulp
in the first and second stages and to pilot refined TMP pulp.
As can be seen in Fig. 22 the proportion of the R14 fibre fraction decreased
from its initial value of 50% to around 12% when a total energy input of 2.4
MWh/t was achieved. The proportion of the R28 fraction did not change as a
function of specific energy input while proportions of the R48, R200 and P200
fraction increased with energy input. The R14 fraction was at a much lower level
than that of the reference pulps using the same specific energy input, while
proportion of the R48 fraction was higher for ESCO pulps. In the cases of other
fibre fractions negligible differences were observed.
The fibre length is presented in Fig. 23 where circle points refer to the
reference mill pulp. Fibre length decreased in ESCO refining as energy input
increased but was at a lower level compared to mill pulp. When comparing fibre
length to a pilot refined pulp (raw material was the same as in pulp pad refining
experiments) the fibre length was at the same level. The freeness (Fig. 24)
decreased as a function of the specific energy input but it remained at a very high
level compared to mill TMP pulps. In the same figure there is a curl index of pulp
pad refined pulps. Pulp pad refining produces highly curled fibres which explain
the measured high freeness values. An additional curl-experiment was performed
to clarify the effect of pulp pad refining on the freeness of pulp.
Thermomechanical pulp with an initial freeness value of 120 ml was refined using
a specific energy input of 0.57 MWh/t. The freeness and curl index of the pulp
before and after this experiment are also presented in Fig. 24. The curl index
increased from 8% to 21% and the freeness from 120 ml to around 330 ml.
Mass proportion of BMN fraction [%]
Fig. 22. Proportions of BMN fractions of pulp pad refined pulps and the first and
second stage mill pulps.
Fibre length [mm]
Fig. 23. Fibre length as a function of specific energy input. Pulp pad refined pulps
mill refined pulps ο and pilot refined pulp .
Curl index [%]
Fig. 24. Curl index and freeness as a function of SEC. Pulp pad refined pulps (initial
freeness of 750 ml)
pulp pad refined pulp (curl experiment)
and mill refined
Microscopic pictures of the fibre fractions (R14, R28, R48 and R200) of coarse
pulp used in experiments, as well as commercial first stage pulp and ESCOrefined (pulp pad refined) pulps are presented in Figs. 25 and 26. Compared to the
microscopic pictures taken before refining, it can be seen that pulp pad refining
effectively develop fibres. ESCO refining produced highly fibrillated R28 and
R48 fractions compared with second stage pulp, but the R200 fraction seems to
have been more fibrillated in the case of the commercial second-stage pulp.
Peeling of fibres was measured using a microscopic “band” method. The
proportion of the inner fibre wall exposed, as presented in Fig. 27, was greater at
higher specific energy inputs and pulp pad refining produced almost totally
peeled fibres. In the same figure three points were measured for the pilot refined
TMP pulps. Fibres of pilot refined pulps were less peeled compared to fibres
refined in pulp pad refining.
Tensile strength was measured for mixed pulp that consisted of pulps refined
with higher specific energy consumptions (circled points in Fig. 28). The result
was a tensile index of 26 kNm/kg which was poor compared to the first (28
kNm/kg) and second (46 kNm/kg) stage pulps. The reason lies behind the high
curl-index because high curl has been reported to affect the strength of pulp. On
the other hand, pulp pad refined pulps contain some untreated coarse pulp and
shives from amongst the plate needles that confuse strength measurements.
R 14 CSF 750 ml
R 28 CSF 750 ml
R 14 2nd stage
R 28 2nd stage
R 14 pulp pad refined
R 28 pulp pad refined
Fig. 25. Scanning electron microscopic (SEM) pictures of BMN +14 and +28 fractions
of coarse TMP pulp with freeness of CSF 750 ml (before pulp pad refining), of mill pulp
(second stage, 2.2 MWh/t) and of pulp pad refined pulp.
R 48 CSF 750 ml
R 200 CSF 750 ml
R 48 2nd stage
R 200 2nd stage
R 48 pulp pad refined
R 200 pulp pad refined
Fig. 26. Microscopic pictures of BMN +48 and +200 fractions of coarse TMP pulp with
freeness of CSF 750 ml (before pulp pad refining), of mill pulp (second stage, 2.2
MWh/t) and of pulp pad refined pulp (~2.2 MWh/t).
Proportion of exposed
inner fibre wall [%]
Fig. 27. Proportion of exposed inner fibre wall as a function of specific energy input of
pulp pad refined pulps and pilot refined TMP pulps (published by permission of
Sabine Heinemann, KCL).
Fig. 28. Pulps from three different refining experiments were mixed together and
strength properties were measured from this pulp mixture. Circled points present the
Power consumption distribution in a refiner disc gap
Measured consistency profiles for first stage and second stage refiners are
presented in Fig. 29 (Papers I and II). Fig. 30 presents consistency profiles for
standard and LE-segments (Paper III). In both figures circular points are
measured values. During calculations it was assumed that consistency increases
monotonously in a refiner disc gap. Due to the curved shape of measured profiles
different monotonously increasing consistency profiles were used as highlighted
in Figs. 29 and 30 by grey areas.
Fig. 29. Measured consistency profiles and grey areas used in calculations for first
stage refiner (left) and second stage refiner (right), presented in Papers I and II.
Fig. 30. Measured consistency profiles and grey areas used in calculations for first
stage refiner with standard (left) and LE-segments (right), presented in Paper III.
The calculated results are presented in Figs. 31 and 32. The cumulative power
distribution in the first and second stage refiners are presented in Fig. 31. In the
first stage refiner more power is consumed in the inner parts of the refiner while
in the second stage refiner the power is consumed mainly in the outer parts. When
LE-segments were used in the first stage refiner, the power distribution had a
similar shape to the second stage refiner, as can be seen in Fig. 32.
Fig. 31. Cumulative power consumption in the disc gap of first and second stage
Fig. 32. Cumulative power consumption in the disc gap of the first stage refiner. LEand standard refiner segments were used.
Development of pulp properties also depends on the segment geometry used.
Freeness as a function of disc radius for standard and LE-segments is presented in
Fig. 33. Freeness decreased more rapidly in outer parts of the refiner equipped
with LE-segment compared to a standard refiner. After first stage refining pulp
properties were measured and are presented in Table 2. The specific energy
consumption was the same in both trials but pulp properties varied: LE-segments
produced shorter fibres and lower freeness.
Fig. 33. Freeness as a function of disc radius for LE- and standard refiner segments.
Table 3. Pulp properties measured after the first stage refiner for Standard and LEsegments.
Fibre length [mm]
BMN +14 fraction [%]
BMN +28 fraction [%]
BMN +100 fraction [%]
BMN +200 fraction [%]
BMN -200 fraction [%]
The aim of the work was to obtain new information about refining mechanisms
by studying the fibre-to-fibre interaction by equipment of shear and compression,
and by calculating energy distribution based on mill trials and theoretical
calculations. In Chapter 5.1 the applicability of equipment for studying refining
phenomenon and results of shearing experiments are discussed. In Chapter 5.2 the
theoretical method to calculate the power distribution and its limitations are
considered. In the Chapter 5.3 the achieved results are pulled together and based
on these results the refining mechanisms are discussed. Plausible reasons for high
energy consumption of refining process are also proposed.
Equipment of shear and compression and fibre-to-fibre refining
In equipment used for shear and compression pulp was heated using steam,
compressed between two plates and then the bottom of the vessel was rotated.
The behaviour of pulp under compressing and shearing forces was measured as a
volume fraction of pulp (Eq. 3) and as a disruptive shear stress (Eq. 4). The aim
was to obtain new information about refining mechanisms, however, compared to
the laboratory equipment, the situation is much more complex in the real refiner.
In the refiner, the rotational speed is very high as fibre and steam flows are
very complex while volume fraction of pulp in the refiner disc gap changes with
time and location. The pulp is also in motion in the refiner disc gap and thus the
disruptive shear stress does not describe the state within a refiner. Heating of pulp
in the equipment of shear and compression was achieved using steam flowing
through the pulp pad and the heating time was so high that pulp was totally heated
and softened. Inside the refiner heating of pulp occurs by frictional forces
(compressing and shearing forces) and also by means of generated steam. The
separation of fibres is strongly dependent on how wood softens and in this respect
the situation within the equipment of shear and compression does not correspond
to the situation in the real refiner. The volume fraction of pulp and disruptive
shear stress are, however, exactly determined magnitudes that can be used to
describe the behaviour of pulp under compressing and shearing forces. These
forces affect pulp in the refiner disc gap and it is thus valuable to understand pulp
behaviour under these forces even more so in a simplified environment.
Results of shearing experiments showed that friction inside the compressed
pulp is dependent on temperature, freeness and fine material. The lower the
temperature the higher the friction inside the pulp was. Removal of fines also
decreased the friction inside the pulp which indicates that fine material works as a
kind of lubricant in shearing. The role of friction inside the refiner is not clearly
understood. However, the energy consumption of refining has been reported to
decrease using refiner plates from which steam and fine material were removed
through perforated plates (Johansson & Richardson 2005). In these plates the
refining zone temperature was much lower compared to standard plates and also
the amount of fines in the refining zone was decreased. The reason for more
efficient refining is probably due to the increased friction between fibres because
of the removal of fines, as well as a reduction in temperature decreases the
friction inside the pulp.
The decreased friction in the refiner can affect refining, however, also in the
desired way. Efficient refining has been reported by using high temperature
refining (Höglund et al. 1997) and the friction in this case is lower compared to a
standard TMP refining. To achieve the desired refining effect (or load), the pulp
has to be compressed more than refining at lower temperatures. The result is a
smaller disc clearance at higher temperatures and thus harsher refining conditions.
The high temperature probably enables the harsh refining without undesirable
The development of fibres under compressing and shearing forces was also
studied. Experiments were done using a temperature of 150 °C and compressive
pressures of 1 and 3 bars. Conditions in the pulp pad refining experiments were
not optimised at all. The results of the experiments showed that pulp pad refining
separates the fibres from each other. In fibre-to-fibre refining the outer layers of
fibres had effectively been peeled off (Fig. 27) while the fibres themselves were
highly fibrillated (Figs. 25 and 26).
A proportion of coarse fibre fraction (R14) decreased to a very low level
when energy input was increased, and a proportion of the middle fibre fractions
and fines increased too (Fig. 22). Comparisons with commercial first and secondstage pulps showed that the amount of coarse fibre fraction (R14) was much
lower in the case of ESCO pulp at the same specific energy consumption while
the amount of the R48 fraction was much higher. Other fibre fractions did not
differ much from those in the commercial TMP pulps. The “band” analysis (Fig.
27) performed to study the peeling of the outer layers of the fibres showed that a
proportion of the inner fibre wall that was exposed increased with specific energy
consumption, and pulp pad refining produced a pulp with more peeled fibres
compared to pilot refined pulp. Microscopic photographs also showed highly
fibrillated fibres (Figs. 25 and 26). The fines content of pulp refined using the
ESCO device was slightly lower than that of the commercial pulps and the
microscopic picture showed a less fibrillated R200 fraction. Pulp pad refining
causes the fibres to fibrillate but seems not to separate the fibrils from the fibre
The freeness of pulp when refined in a pulp pad was very high (Fig. 24),
much higher than that of a commercial TMP pulp, even though other fibre
properties were fairly similar. The reason lies in the very high curl index of the
former, as presented in Fig. 24, due to the fact that latency of pulp reduces its
filtration resistance (Chapter 2.2.3). To verify this result, a pulp pad refining
experiment was performed using pulp with an initial freeness value of 120 ml,
whereupon the freeness increased to a much higher level (330 ml) during ESCO
refining. At the same time, the curl index increased from its initial level of 10% to
around 21%. High curl index after pulp pad refining may derive from the
experimental arrangement – pulp was cooled off in a curled state in the ESCOequipment and did not re-straighten in hot-disintegration. Otherwise, inside the
refiner, fibre-to-segment contacts in the outer refining zone (Zone III in Fig. 1)
probably straighten fibres. Due to high curl index of pulp the strength properties
of pulp pad refined pulps were very much lower compared to the commercial
Pulp pad refining produced pulp with generally similar papermaking
properties to commercial TMP pulps. It effectively fibrillated and peeled the
fibres, and coarse fibre fraction decreased rapidly to a very low level. It seems
that contacts between fibres are also important for fibre development inside a
commercial TMP refiner.
The method to calculate power consumption distribution
The method to calculate power consumption distribution was based on mass and
energy balances and also on mill trials in which consistency profile and
temperature profile, as well as production rate was measured. For calculations, the
following assumptions and simplifications were made and all the possible
inaccuracies in the calculation were derived from the following sources:
Steam was assumed to be saturated in a disc gap. Fibre saturation point (the
point at which all water is bounded in the cell walls) generally falls between a
moisture content of 25 to 30% (Panshin & De Zeeuw 1964) which means that
up to 70% consistency, there is still free water. Thus the assumption of
saturated steam is justified in the refining consistency range.
Enthalpy of fibres was assumed to be in accordance to an equation
determined for wood. There are no available enthalpy values determined for
spruce TMP pulps. The enthalpy of wood is used to calculate the energy
needed to heat pulp inside the refiner. This energy is very low compared to
the total refining energy and if there is a slight inaccuracy in the enthalpy
values of pulp its effect on the result of the calculation is insignificant.
Fibres, water and steam were assumed to have equal temperatures. The
temperature in a disc gap is measured by a PT-100 temperature sensor. The
steam, water and fibre temperatures are not measured separately and their
local temperatures can be different from those measured by the sensor. The
effect of a temperature profile on the result of power consumption calculation
is not, however, the critical factor. Changing the maximum temperature by
10% means the total power consumption changes less than 1% (Paper I).
No steam flow was assumed to be over the point of maximum temperature. In
refining there are steam flows in both directions (blow-back steam and
forward steam). It is generally assumed that there is no steam flow over the
point of maximum temperature (e.g., Miles & May 1990) and also in the
calculations presented here. In the Paper I the assumption was not applied,
but the results were comparable to the results of Papers II and III. The
assumption thus is not a critical factor for calculated results.
Consistency profile was assumed to be in accordance to measured values, the
feed consistency was assumed to be at a certain level. Sampling from the
refiner plate gap is very challenging (sampling process is presented in Papers
I, II and III) and the reliability of the measured consistency profiles can be
questioned. The shape of the consistency profile has been proposed due to the
dilution water not being totally mixed with the pulp (Paper II). It was
assumed that water can mix easily with pulp in the second stage refiner
compared to chips in the first stage refiner. An acceptable reason for the
strange shape found in the consistency profile can be due to back-flow of
fibres in the stator side. The drier pulp from the outer parts of the refiner is
flowing back in the stator side and the sample is thus not a representative
sample of the whole pulp.
In order to take account of the uncertainties connected to the consistency profile
calculations were done using different monotonously increasing consistency
profiles or grey areas in Figs. 29 and 30.
Results of calculations showed that power distribution depends on segment
geometry used in a refiner, as well as the refiner stage. In the first stage refiner
more power was consumed in the inner parts of the refiner compared to the
second stage refiner. The breaking of chips occurs in the inner area of the first
stage refiner and there is an intensive mixing and backflow of fibres. In the
second stage refiner mechanisms are not similar because the feed of the second
stage refiner is pulp instead of chips. Due to the different mechanisms the power
distribution is also different.
When using unidirectional LE-segments the mixing and pulp residence time
in the inner parts (Härkönen et al. 1999) as well as energy consumption in this
area are reduced, compared to standard refiner segments. Differences in the pulp
properties with these different segment types were, however, noteworthy. The role
of the inner refining zone and its high energy consumption seems to be important
for final pulp quality as well.
Mechanisms of refining – Why energy is dissipated?
It is possible to distinguish three zones (Fig. 1) inside the refiner. The role of
these zones on energy consumption of refining, separation and development of
fibres seems to be different. There is a chip breaking zone at the centre of the
refiner where chips are broken down rapidly into coarse pulp. In the inner parts of
the refiner the disc gap is high and quantity of pulp has been shown to be high too.
Consequently inter-fibre contacts play an important role. In the outer parts of the
refiner the segment area is higher, disc gap narrower and the coverage of pulp is
lower compared to inner parts. Thus contacts between segment surface and pulp
become more important in the outer parts of the refiner.
In this work inter-fibre refining or pulp pad refining was studied. The
equipment of shear and compression aims to explain the phenomena mainly
located in the inner refining zone of the refiner. The fibre-to-segment contacts
were not considered in this work. However, based on pulp pad refining
experiments it can be assumed that the function of the outer refining zone is to
separate fibrils from fibre surfaces, and possibly straighten curled fibres.
The results of this study (Papers I, II and III) showed that with standard
refiner segments power consumption was considerable in the inner area of disc
gap (zones I and II, Fig. 1) where inter-fibre contacts play an important role. In
this work inter-fibre refining or pulp pad refining was studied in Papers III and IV.
Pulp pad refining (Paper V) produced pulp with generally similar fibre
properties to commercial TMP pulp (Chapter 5.1), even though conditions in
experiments were not optimised. The coarse pre-refined pulp (720 ml) used as an
initial material in pulp pad refining experiments had been produced in a pilot
refiner with a specific energy consumption of 0.6 MWh/t. When this material was
submitted to pulp pad refining at a specific energy level of 0.6 MWh/t with a total
specific energy of 1.2 MWh/t, which is comparable to the first stage in a normal
mill, fibre properties were also comparable. Energy required to separate fibres has
been reported to be insignificant compared to energy needed to fibrillate fibres
(Uhmeier and Salmén 1996). However, energy needed to break chips into shives
in pilot refining was equal to energy required to develop shives into fibrillated
fibres in pulp pad refining. The cause of high energy consumption of today’s TMP
process probably derives from the phenomena in the inner parts of the refiner disc
gap where refining appears not to occur in the most efficient way.
The factors affecting the pulp pad refining are temperature, compressive force
and friction between pulp and segment surface. Temperature is the key factor
affecting fibre separation, and it has to be high enough to soften fibres and to
prevent fibre cutting. The compressive force affects the internal friction in the
pulp: the higher the compressive force, a more compressed state the pulp is and
higher is the frictional force inside the pulp pad. At the moment it is not known
what optimum force inside the pulp pad is required to achieve the most effective
fibre development. Too low a compressive force will mainly cause heating of the
pulp by frictional forces without any changes in fibre structure, while too high a
compressive force may cause undesirable damage to fibres. In the inner parts of
the refiner, before the narrow disc gap the compressive force acting on the pulp is
probably too low to achieve effective refining between fibres. The existence of a
too low compressive force results in not enough friction between the fibres, so
that the pulp is heated without any changes in fibre structure. This is most
probably the basic cause of high energy consumption in the operation of current
On the other hand, heating of pulp is extremely important and the role of
inner refining zone is probably to heat wood material appropriately in order to
separate fibres in a optimal way. Too low a temperature results in broken fibres
and a similar effect can be observed if the residence time of pulp in the inner parts
of the refiner is too low. Power distribution calculations showed decreased power
in the inner parts of the refiner but also poorer pulp properties. In one situation
too high a temperature affects lignin covered fibres that can not be refined any
further. However, the energy needed to heat fibres is minimal compared to
refining energy while the heating of fibres could certainly be preformed more
The results of pulp pad refining experiments showed that repeated
compression and relaxation actions due to grooved discs are not required in order
to further develop fibres that are suitable for papermaking; even if it has been
mentioned to be an important mechanism inside a refiner. The function of a
refiner segment surface is to cause a high tangential frictional force that such
energy is efficiently transferred into the pulp pad. If friction is too low there will
be no pulp pad refining at all, but energy consumption will occur at the gliding
interface between pulp and segment surfaces. Results of the shearing experiments
with smooth surfaces and bar type segments (Fig. 13) showed that even with the
use of real segments it did not give a perfect energy transfer inside the pulp but a
gliding interface between pulp and the segment surface occured. This suggests
that the tangential frictional force between segment surface and the pulp in a
commercial TMP refiner is not high enough to transfer energy entirely into the
pulp pad but that some gliding takes place between pulp and the segment surface.
Some energy inside the refiner is probably dissipated due to the frictional force
between segment surface and the pulp being too low.
To attain an effective pulp pad refining inside the refiner a highly compressed
pad is required. This calls for large amounts of pulp in the disc gap and the
segment surface should be such that energy is very effectively transferred into the
pulp. The pulp should be in such a compressed state that its shearing causes high
enough frictional forces inside the pulp pad to cause efficient development of
fibres and not only to heat them.
Results presented in this thesis provide new information on refining mechanisms
and especially on the phenomena in the inner refining zone. Based on studies
done in Papers I, II and III, hypothesis one has be stated to be true. In the first
stage refiner with standard refiner segments, the power consumption seems to be
considerable in the inner area of disc gap where contacts between fibres are
important. Based on studies presented in Papers IV and V, hypothesis two has also
been stated to be true. Effective refining can be achieved in fibre-to-fibre refining
without any impacts due to bar crossing if friction between segment surface and
pulp is high enough and if pulp is in such a compressed state that pulp pad
refining can occur. The high energy consumption of thermomechanical pulping
and ineffectiveness of the refining process probably derives from having a too
low frictional force that heats pulp and evaporates water without any changes in
the fibre structure.
The method to calculate power distribution in a refiner disc gap gives a new
tool for improving the present refining process and developing new energy saving
refiner segments. Results of efficiency of the fibre-to-fibre refining may also give
ideas for developing today’s refiners – how to achieve the effective pulp pad
refining when effectiveness of pulp pad refining depends on, e.g., temperature
and freeness of pulp (Paper IV). Otherwise, the results of this study may give
ideas for developing totally new energy saving mechanical pulping processes.